INVESTIGADORES
PONZONI Ignacio
congresos y reuniones científicas
Título:
Constrained Partitioning of Sparse Matrices Using Graph Decompositions
Autor/es:
PONZONI, IGNACIO; SÁNCHEZ, MABEL C.; BRIGNOLE, NÉLIDA B.
Lugar:
Toronto, Canadá
Reunión:
Conferencia; SIAM Annual Meeting 1998; 1998
Institución organizadora:
SIAM (Society for Industrial and Applied Mathematics)
Resumen:
A direct method based on graph decompositions that carries out structural rearrangements of huge sparse occurrence matrices, also comprising structurally singular matrices, was developed. The procedure yields a Block lower-Triangular Form specially adapted to solve plant instrumentation problems. The decomposition is obtained through a new vertex classification. Another innovative feature is the incorporation of partitioning constraints to ensure realistic results. The final decomposition is optimum because it yields the maximum number of irreducible blocks.